Forest Ecology and Management 285 (2012) 1–10

Contents lists available at SciVerse ScienceDirect

Forest Ecology and Management

journal homepage: www.elsevier .com/locate / foreco

Drought-induced embolism in current-year shoots of two Mediterraneanevergreen oaks

Clara A. Pinto a,⇑, Jorge S. David b,d, Hervé Cochard c, Maria C. Caldeira b,d, Manuel O. Henriques b,d,Teresa Quilhó e,d, Teresa A. Paço b,f, João S. Pereira b,d, Teresa S. David a,d

a Instituto Nacional de Investigação Agrária e Veterinária, I.P., Quinta do Marquês, Av. da República, 2780-159 Oeiras, Portugalb Instituto Superior de Agronomia, Technical University of Lisbon, Tapada da Ajuda, 1349-017 Lisboa, Portugalc INRA, UMR 547 PIAF, F-63100 Clermont-Ferrand, Franced Centro de Estudos Florestais, Tapada da Ajuda, 1349-017 Lisboa, Portugale Centro das Florestas e Produtos Florestais, Instituto de Investigação Científica Tropical, Tapada da Ajuda, 1349-017 Lisboa, Portugalf Centro de Engenharia dos Biossistemas, Tapada da Ajuda, 1349-017 Lisboa, Portugal

a r t i c l e i n f o a b s t r a c t

Article history:Received 14 April 2012Received in revised form 3 August 2012Accepted 4 August 2012Available online 1 September 2012

Keywords:Quercus suberQuercus ilexVulnerability curvesHydraulic conductivityXylem anatomyHydraulic safety margins

0378-1127/$ - see front matter � 2012 Elsevier B.V. Ahttp://dx.doi.org/10.1016/j.foreco.2012.08.005

⇑ Corresponding author. Tel.: +351 214463798; faxE-mail addresses: clara.pinto@iniav.pt (C.A. Pinto

David), herve.cochard@clermont.inra.fr (H. CochardCaldeira), mhenriques@isa.utl.pt (M.O. Henriques), tertapaco@isa.utl.pt (T.A. Paço), jspereira@isa.utl.pt (J.S. P(T.S. David).

Quercus suber and Quercus ilex trees are major elements of Mediterranean landscapes, which are threa-tened by the increasing water deficits related to climate change. To contribute to the understanding ofthe capacity of these oaks to withstand severe drought we assessed the vulnerability to xylem embolismand the anatomical traits in current-year shoots. Data were collected in mature trees at two sites, in cen-tral/coastal and southern/inland Portugal. In situ safety margins to hydraulic failure were evaluated fromlong-term predawn and midday leaf water potential records. Results showed that xylem vulnerability toembolism was similar in Q. ilex and Q. suber. The 50% loss in hydraulic conductivity (Wxyl,50PLC) wasobserved at xylem water potentials of �2.9 and �3.2 MPa in shoots of Q. suber and Q. ilex, respectively.Values of mean vessel diameter of Q. suber shoots at both sites suggest an intra-species adaptation tothe local water availability, with larger vessels at the more mesic site. In situ hydraulic safety marginsobserved in shoots showed that, even during the driest periods, both oaks lived comfortably above themost critical embolism thresholds. However, the hydraulic safety margins were narrower in the driestsite. Results are relevant to the understanding of survival, growth, and functional behaviour of evergreenoaks in Mediterranean climates, under recurrent/seasonal drought conditions.

� 2012 Elsevier B.V. All rights reserved.

1. Introduction

In recent decades there were many reports of enhanced treemortality rates in forests of all ages, especially in temperate re-gions (van Mantgem et al., 2009), but also in Mediterranean-typeclimates (Lloret et al., 2004). Plausible causes are increasing waterdeficits related to global warming (van Mantgem et al., 2009),which may result in the failure of the hydraulic system of trees(Ryan, 2011), associated or not to starvation and depletion of car-bon reserves (McDowell et al., 2011). During dry periods waterstress may induce xylem cavitation and the formation of embo-lisms, resulting in the disruption of water columns and in thereduction of water supply to leaves (Tyree and Zimmermann,2002). Xylem embolism caused by severe drought stress has been

ll rights reserved.

: +351 214463702.), soaresdavid@isa.utl.pt (J.S.), mcaldeira@isa.utl.pt (M.C.isantos@isa.utl.pt (T. Quilhó),ereira), teresa.david@iniav.pt

considered one of the major factors affecting plant productivityand survival (Tyree and Sperry, 1988). Mediterranean-climate re-gions are characterised by recurrent droughts, with irregular/lim-ited rainfall and high evaporative demand. Climate changescenarios for the Western Mediterranean Basin foresee warmerair temperatures and an increase in the length and intensity ofthe seasonal summer drought (Miranda et al., 2002). Trees copewith these seasonal water shortages by preventing water lossesthrough stomatal closure and maximising the soil and groundwa-ter uptake by deep roots (Canadell et al., 1996; David et al.,2007; Maherali et al., 2004). Under extreme drought, stomatal reg-ulation may not be enough to maintain leaf water potential above acritical threshold and catastrophic embolism may occur (Sperry,1986).

Mediterranean evergreen oak woodlands (montados in Portugaland dehesas in Spain) characterise the landscape of extensive areasof the Iberian Peninsula and are ecosystems of a high socioeco-nomic and conservation value (Bugalho et al., 2011). In Central-Southern Portugal, Quercus suber L. (cork oak) and Quercus ilexspp. rotundifolia Lam. syn. ballota (holm oak) are the dominant

Table 1Mean (standard deviation, SD) morphometric data of the four sampled trees (Q. suberand Q. ilex) at the experimental sites.

Site Species DBH (m) Height (m) Crown projectedarea (m2)

Lezirias Q. suber 0.73 (0.18) 12.82 (1.16) 208.1 (32.4)Mitra Q. suber 0.49 (0.04) 9.05 (0.44) 128.7 (25.8)Mitra Q. ilex 0.40 (0.05) 7.63 (0.48) 60.2 (20.2)

DBH is diameter at breast height.

2 C.A. Pinto et al. / Forest Ecology and Management 285 (2012) 1–10

species. Although co-occurring in some places, Q. ilex prevails inthe inner, drier regions and Q. suber in the wetter, coastal and mon-tane areas. This geographic distribution, though influenced by hu-man activity, seems to mainly reflect the lesser drought resistanceof Q. suber (David et al., 2007). Extensive literature is available con-cerning the xylem vulnerability to embolism and other hydraulictraits of Q. ilex (Corcuera et al., 2004; Gartner et al., 2003; Limousinet al., 2010; Lo Gullo and Salleo, 1993; Martínez-Vilalta et al., 2002;Tognetti et al., 1998; Villar-Salvador et al., 1997) and of other Med-iterranean species (Iovi et al., 2009). However, Q. suber has beenmuch less studied (Lo Gullo et al., 2003; Tyree and Cochard,1996; Vaz et al., 2012). To our knowledge, its vulnerability todrought-induced embolism has not yet been fully characterisedparticularly in adult trees.

We aimed to evaluate and compare the xylem conducting effi-ciency and vulnerability of current-year shoots of Q. suber and Q.ilex to drought induced cavitation. Measurements and samplingwere done in mature trees at two different sites, located in Central(Lezirias) and Southern (Mitra) Portugal. Cork oak was studied atLezirias and Mitra whereas holm oak was only studied at Mitra.We hypothesised that: (1) Q. ilex xylem would be less prone tohydraulic failure than Q. suber, and the difference would be enoughto explain the geographic distribution of the two species; (2) Dif-ferences in site water availability might reflect in small intra-spe-cies variations in the vulnerability to embolism and xylemanatomy. The specific objectives of this work were to: (1) measurexylem vulnerability to embolism in current-year shoots of maturecork and holm oak trees; (2) test intra-species differences in vul-nerability to embolism and xylem anatomy of cork oak shoots be-tween sites; (3) evaluate if species are living close to theirhydraulic limits, by estimating the in situ minimum safety margins.

2. Material and methods

2.1. Study sites and plant material

The study was carried out at two evergreen oak woodlands90 km apart, in Central (Lezirias site) and Southern (Mitra site) Por-tugal. The Mitra site (38�320N, 8�000W) has an inland location nearthe town of Évora, at the ‘‘Herdade da Alfarrobeira’’ (ca. 150 kmSouth-East of Lisbon), in a sparse mixed stand where Q. ilex domi-nates and Q. suber occurs in scattered patches. At this site two sin-gle species plots (150 m apart) were established. The soil is a 1 mdeep Dystric Cambisol (FAO, 1988), with low water retentioncapacity, overlying a granite bedrock. The Lezirias site (38�500N,8�490W) is located near the coast, at the estate of ‘‘Companhiadas Lezirias’’, about 50 km East of Lisbon, in a pure Q. suber stand.The soil is a deep Arenosol (FAO, 1988), with high permeability andlow water retention capacity, overlying a thick clay layer at ca. 9 mdepth. The climate is Mediterranean at both sites, with hot drysummers and wet mild winters. Rainfall occurs predominantlyfrom October to April. Long-term (1951–1980) mean annual rain-fall and open water evaporation are 665 and 1760 mm for Mitra(the more xeric site), and 708 and 1347 mm for Lezirias, respec-tively (INMG, 1991a, 1991b). Mean annual temperature is 15.0 �Cat Mitra, ranging from 8.6 �C in January to 23.1 �C in August, and15.6 �C at Lezirias, ranging from 9.9 �C in January to 22.0 �C in Au-gust (INMG, 1991a, 1991b).

Hydraulic and anatomy measurements were done in 2007–2008 on four mature trees per species and site. Mean morphomet-ric data of the sampled trees are given in Table 1. Tree age is about60 years at Lezirias and 70 (Q. ilex) – 80 (Q. suber) years at Mitra.Average annual growth of current-year shoots was 5 cm and 3–4 cm for Q. ilex and Q. suber at Mitra, respectively, and 7–20 cmfor Q. suber at Lezirias (Pinto et al., 2011). Branches of Q. suber weresampled from both sites, whereas branches of Q. ilex trees were

only sampled from the Mitra site. In situ leaf and xylem water po-tential measurements were done in the same trees used for branchsampling.

2.2. Environmental variables

Solar radiation (CM6B, Kipp and Zonen, Delft, The Netherlands)and rainfall (tipping-bucket rain gauge recorder ARG100, Environ-mental Measurements, Gateshead, UK) were measured at bothsites (2001–2003 at Mitra and 2006–2008 at Lezirias). Water tabledepth was also measured at both sites by pressure transducers(PDCR 830, Campbell Scientific) installed in boreholes. Data wererecorded every 10 s and stored as 10 or 30 min averages or totalsby CR10X data loggers (Campbell Scientific, Shepshed, UK).

2.3. ‘‘In situ’’ leaf water potentials

Leaf water potential (Wl, MPa) was measured monthly at Mitra(2001–2003 for both plots) and Lezirias (2006–2008). Measure-ments were done in four mature trees per plot at predawn (Wl,pd)and around midday (Wl,md), using a Scholander pressure chamber(PMS 1000, PMS Instruments, Corvalis, Oregon, USA) (Scholanderet al., 1965). At each sampling time, three to four leaves per treewere collected at similar heights from the South-facing part ofthe crowns, bagged, and immediately measured.

2.4. Vulnerability to xylem embolism

The degree of vulnerability to embolism was inferred from vul-nerability curves (VCs), plotting the xylem water potential (Wxyl,MPa) versus the corresponding percentage loss of hydraulic con-ductivity (PLC, %). We used the dehydration technique (Sperry,1986; Sperry et al., 1988), considered as the most reliable (refer-ence technique) (Choat et al., 2010; Cochard et al., 2005; Sperryet al., 2012).

The hydraulic conductivity (Kh, kg s�1 MPa�1 m) of shoot seg-ments of both species was measured following Sperry et al.(1988), with a high precision flow meter, XYL’EM (Embolism Me-ter, Bronkhorst, Montigny-Les-Cormeilles, France). Kh was mea-sured at low pressure (2–3 � 10�3 MPa), to minimise thedisplacement of air bubbles in open vessels. All segments wereperfused with ultra-pure, deionized, degassed and filtered(0.2 lm) water with 10 mM KCl. Kh was calculated as the ratio be-tween the flow through each segment and the correspondinghydrostatic pressure gradient. It was converted to specific hydrau-lic conductivity (Ks, kg s�1 MPa�1 m�1) by dividing by the xylemcross-sectional area (m2).

Prior to branch collection, the maximum length of xylem ves-sels was determined using the air infiltration method in the entirelength of large branches (Zimmermann and Jeje, 1981). The longestvessels in Q. suber and Q. ilex stems were found to be 2.6 and 2.0 m,respectively. Hence, to avoid contamination by air entry upon cut-ting, only branches longer than 3 m were sampled, overnight, fromthe top third of the South-facing side of the crowns. At each sam-pling date, one or two large branches per tree (depending on crown

C.A. Pinto et al. / Forest Ecology and Management 285 (2012) 1–10 3

size) were harvested and immediately bagged to minimise dehy-dration. Once in the laboratory, branches were gradually dehy-drated to cover a wide range of Wxyl values. For each Wxyl point(measured in previously covered leaves), PLC values were obtainedfrom 3 to 6 terminal, and non-contiguous current-year shoots. Onaverage, 20 current-year shoots were sampled per branch. Aftermeasuring Wxyl, shoots with a diameter of 2–3 mm were excisedunderwater, trimmed to approximately 2.5 cm length, and the ini-tial hydraulic conductivity (Kh,in) was measured. Segments wererepeatedly perfused (pressure of ca. 0.15 MPa) to dissolve and ex-pel air bubbles, until the maximum conductivity (Kh,max) was ob-tained. The degree of embolism (PLC, %) was estimated based onKh,in and Kh,max (Sperry et al., 1988). Only current-year shoots wereused for conductivity measurements to prevent previously embol-ized and non-functional conduits of older xylem rings from con-tributing to PLC estimates. Sample size was restrained by theaverage length of current-year shoots of Q. suber at Mitra. Quercusspecies are known to have very long vessels and, hence, it is likelythat some samples had open-cut vessels at both ends. However,sample length should not impose a methodological problem to vul-nerability estimates (i.e., overestimation) since PLC values are ra-tios of conductivities.

Vulnerability curves (VCs) were fitted using the exponential sig-moid equation proposed by Pammenter and Vander Willigen(1998):

PLC ¼ 1001þ eaðWxyl�bÞ ð1Þ

where Wxyl (MPa) is the measured xylem water potential; a is re-lated to the slope of the curve and b is the Wxyl at which a 50% lossin hydraulic conductivity occurs (Wxyl,50PLC). Although there areother functional options to fit the VCs (Ogle et al., 2009) we decidedto use the Pammenter and Vander Willigen (1998) equation since itis simple, widely used, and parameters have an inherent physiolog-ical meaning (under the air-seeding hypothesis, Tyree and Zimmer-mann (2002)).

Three vulnerability curves were constructed: two for Q. suber,one for each site, and one for Q. ilex, for the Mitra site. Each VCwas adjusted using at least 25 points per species/site.

A modified version of Eq. (1) was also fitted to the PLC/Wxyl

dataset to account for a non-zero ‘‘native’’ level of embolism(PLCnat):

PLC ¼ PLCnat þ100� PLCnat

1þ ea1ðWxyl�b1Þð2Þ

Wxyl,50PLC values obtained from Eqs. (1) and (2) were used to assessthe possible impact of PLCnat on vulnerability estimates.

The xylem tension/pressure values causing 50% (Wxyl,50PLC), 75%(Wxyl,75PLC) and complete loss of conductivity (Wxyl,99PLC) were usedto establish the hydraulic safety margins.

2.5. Xylem anatomy

Vessel inner diameter was measured on a sub-sample of thesegments used for the vulnerability curves (three segments pertree). Transverse sections (ca. 20 lm thick) were obtained using asledge microtome (Leica SM2400, Leica Microsystems GmbH,Nussloch, Germany), stained with safranine (1%) for 1–3 min, andmounted with glycerine Kaiser. Cross sections were viewed in alight microscope system (Leica CTR Mic) interfaced with a digitalcamera (Leica DM LA, Leica Microsystems, Heerbrugg, Switzerland)at 10 � 10 � 2. Two sectors, located 90–180� apart, from the outer-most rings of each transverse section were sampled. Within eachsector, the cross-sectional inner diameters of all open vessels(lm) were measured using the image analysis software LeicaQWin Plus V. 3.5.0 (Leica Microsystems, Heerbrugg, Switzerland).

The following variables were considered to describe xylem anat-omy: mean vessel diameter (Dmean, lm), vessel density (numberof vessels mm�2), and conductive area (percentage of cross-sec-tional area occupied by vessels). A theoretical hydraulic conductiv-ity, Kh,t (kg s�1 MPa�1 m) was calculated for each section, from theanatomy data, following the Hagen–Poiseuille equation (Sellinet al., 2008; Sperry et al., 1994; Tyree et al., 1994; Tyree andZimmermann, 2002):

Kh;t ¼ ðpq=128gÞXn

i¼1

D4i ð3Þ

where q is density of water (1000 kg m�3), g is the dynamic viscos-ity of water (10�9 MPa s at 20 �C) and D is the diameter of the i ves-sel (m). To characterise the xylem conducting efficiency per unit ofsapwood area, the theoretical specific hydraulic conductivity, Ks,t

(kg s�1 MPa�1 m�1), was determined by dividing the estimatedKh,t by the sample cross-sectional xylem area (m2). We also calcu-lated the vessel diameter frequency in 5 lm diameter classes andestimated the relative contribution of each diameter class to thepredicted Kh,t of each segment (Tyree et al., 1994). A hydraulicallyweighted mean diameter (Dhw, lm) was estimated for each seg-ment as: ð

PD5

i =P

D4i Þ (Sperry et al., 1994). The mean diameter of

the vessels that account for 95% of the predicted Kh,t, D95, was alsoestimated (Tyree et al., 1994).

2.6. Estimation of shoot xylem water potential from ‘‘in situ’’ Wl

measurements

Daily courses of leaf and shoot xylem water potentials weredone in two sunny summer days per plot. Measurements (fourtrees per species/site) started at dawn and were repeated every2–3 h throughout the day. At each measurement time, three cov-ered (Wxyl) and three sun exposed (Wl) leaves were measured.

As the number of direct shoot Wxyl measurements was re-stricted to two days per plot, we attempted to find a relationshipto predict Wxyl from the available long-term Wl records. A linearregression between shoot xylem and leaf water potentials wasfound to be site-specific. Using the Darcy’s law (e.g. Wullschlegeret al., 1998) we developed a relationship between Wxyl and Wl,pd

and Wl to tackle the mentioned difficulty. Water flow (F) from soilor groundwater to the leaves can be described as:

F ¼ Wl;pd �Wl

Rplantð4Þ

where Rplant is the total resistance in the soil to leaf pathway andWl,pd is assumed as a surrogate of Wsoil. Considering only the flowpath from the stem xylem to leaves, Eq. (4) may be re-written as:

F ¼ Wxyl �Wl

Rpetioleð5Þ

where Rpetiole is the total resistance in the xylem-leaf pathway.Assuming steady-state conditions in the soil-leaf continuum, Eqs.(4) and (5) can be combined as:

F ¼ Wl;pd �Wl

Rplant¼ Wxyl �Wl

Rpetioleð6Þ

Eq. (6) can be re-written as:

Wxyl ¼ ðWl;pd �WlÞRpetiole

RplantþWl ð7Þ

Considering Rpetiole

Rplantas an approximate constant, a, Eq. (7) can be re-ar-

ranged as:

Wxyl

Wl;pd¼ aþ ð1� aÞ Wl

Wl;pdð8Þ

4 C.A. Pinto et al. / Forest Ecology and Management 285 (2012) 1–10

Data from both oaks fitted nicely to the linear relationship betweenWxyl=Wl;pd and Wl=Wl;pd (Eq. (8)), irrespective of the site. Hence, wecould predict minimum Wxyl of shoots under field conditions fromthe long-term Wl,pd and Wl,md records. These values were then en-tered in the fitted vulnerability curves to establish the minimumhydraulic safety margins at which shoots of both species operate(Hacke et al., 2000; Pockman and Sperry, 2000).

2.7. Statistical analyses

The coefficients a and b of the fitted vulnerability curves wereestimated using the non-linear regression procedure in the statis-tics software package PASW Statistic 18.0.0 (IBM Corporation, NY,USA) and the software package R 2.13.0 (The R Foundation for Sta-tistical Computing, Vienna, Austria) using the ‘‘nls’’ function. Thefitted coefficients a and b were compared between species/sitesand tested with Student t-tests (Snedecor and Cochran, 1978;Zar, 1996).

Differences in mean vessel diameter, vessel density, conductivearea and Ks, between trees, species and sites, were tested withAnalyses of Variance. Linear mixed models were used, consideringspecies/sites as fixed effects and tree within species as a randomfactor. These analyses were performed with the mixed procedureof SAS statistic software (SAS system 9.2, SAS Institute Inc., Cary,NC, USA) for a significance level of 5% (a = 0.05).

3. Results

3.1. Environmental variables

Fig. 1 shows the seasonal variation of solar radiation(MJ m�2 month�1) and rainfall (mm month�1) for the Lezirias andMitra sites. During summer periods, radiation was high and rainfallusually scarce. Annual rainfall at Lezirias was 802.0, 441.5 and

0

200

400

600

800

1000

1200

1400

Jan

Mar

May Ju

lSe

pN

ov Jan

Mar

May Ju

l

2001-20

0

200

400

600

800

1000

1200

1400

Jan

Mar

May Ju

lSe

tN

ov Jan

Mar

May Ju

l

2006-20

Sola

r rad

iatio

n (M

Jm

-2m

onth

-1)

(a) Lezirias

(b) Mitra

0

Jan

Mar

May Ju

lSe

pN

ov Jan

Mar

May Ju

l

Fig. 1. Seasonal variation of total solar radiation (open bars) and rainfall (closed bars) du(2006–2008); (b) Mitra site (2001–2003).

576.0 mm, i.e. 13% above average in 2006 and 38% and 19% belowaverage in 2007 and 2008, respectively. At Mitra, annual rainfallwas 843.0, 825.2 and 706.4 mm, i.e. 27%, 24% and 6% above averagein the years between 2001 and 2003, respectively. Differences inrainfall distribution and amount reflect the variability typical ofMediterranean-type climates.

Water table depth varied from a minimum in winter/spring to amaximum in summer/autumn. During the studied years, it fluctu-ated between 1.3 and 4.5 m at Lezirias, and between 1.1 and 5.1 mat Mitra.

3.2. ‘‘In situ’’ leaf water potentials

Long-term variation in Q. suber Wl,pd differed between sites. AtLezirias, Wl,pd remained high (above �0.45 MPa) (Fig. 2a), with noseasonal variation. At Mitra, where Q. suber and Q. ilex share thesame habitat conditions, Wl,pd remained high during the wet peri-ods, but strongly declined during the dry summer up to �2.3 MPain Q. suber and �1.7 MPa in Q. ilex (Fig. 2b and c). At both sites,Wl,md values varied according to atmospheric conditions. Absoluteminimum for each species was about constant throughout the sea-sons and years and was lower in Q. ilex (�3.48 MPa) (Fig. 2c) thanin Q. suber (�2.98 and �2.95 MPa at Lezirias and Mitra, respec-tively) (Fig. 2a and b).

3.3. Vulnerability to xylem embolism and maximum hydraulicconductivity

The parameters of the VCs fitted through Eq. (1) are shown inTable 2 and Fig. 3. The xylem tension inducing 50% loss of hydraulicconductivity (Wxyl,50PLC; parameter b) was �2.9 MPa for Q. suber(Table 2, Fig. 3a and b) and �3.2 MPa for Q. ilex (Table 2, Fig. 3c).Differences between parameters a and b for Q. suber and Q. ilexwere statistically non-significant, either between species or sites(Table 2). The original VCs (Eq. (1)) showed ‘‘native’’ embolism

0

50

100

150

200

250

300

Sep

Nov Jan

Mar

Ma y Ju

lSe

pN

ov

03

0

50

100

150

200

250

300

Sep

Nov Jan

Mar

Ma y Ju

lSe

pN

ov

08

Rai

nfal

l (m

m m

onth

-1)

300

Sep

Nov Jan

Mar

Ma y Ju

lSe

pN

ov

ring the years with leaf water potential measurements at each site: (a) Lezirias site

-4.0

-3.0

-2.0

-1.0

0.0

-4.0

-3.0

-2.0

-1.0

0.0

Ψl(

MPa

)(a)

(c)

(b)

Q. suber - Lezirias 2006-2008

Q. suber - Mitra2001-2003

Q. ilex - Mitra2001 2003

-4.0

-3.0

-2.0

-1.0

0.0

Jul Nov Mar Jul Nov Mar Jul Nov

Jul Nov Mar Jul Nov Mar Jul Nov

Jul Nov Mar Jul Nov Mar Jul Nov

(c) Q. ilex - Mitra2001-2003

Fig. 2. Long-term seasonal variation of predawn (open symbols) and midday(closed symbols) leaf water potential in Q. suber (a and b) and Q. ilex (c) from theLezirias (2006–2008) and Mitra (2001–2003) sites. Three to four leaves per treewere measured in four trees per species/site. Error bars are standard errors(contained within the size of the symbols in predawn leaf water potential).

Table 2Hydraulic parameters (SD) of current-year shoots of Q. suber and Q. ilex, from the Lezirias anWxyl,50PLC, between species/sites. Each fitted VC was adjusted using at least 25 points permodels considering species/sites as fixed effects and tree within species as a random factor.within species/site were statistically non-significant. Statistics are not given for Wxyl,75PLC

Q. suber

Lezirias MitraParameters (A) (B)

Wxyl,50PLC (MPa) �2.89 (2.0) �2.94 (1.6)a 0.50 (0.07) 0.47 (0.05)Ks (kg s�1 MPa�1 m�1) 1.03 (0.56) 1.03 (0.31)Wxyl,75PLC (MPa) �5.09 �5.28Wxyl,99PLC (MPa) �12.08 �12.72

ns: non-significant.

C.A. Pinto et al. / Forest Ecology and Management 285 (2012) 1–10 5

levels of around 20% PLC. To account for the impact of a non-zero‘‘native’’ level of embolism, a new VC (Eq. (2)) was fitted to eachdataset. Wxyl,50PLC values for the two fitted curves were similar inall cases (Fig. 3).

Mean values of the measured maximum specific hydraulic con-ductivity (Ks) were 1.38 and 1.03 kg s�1 MPa�1 m�1 in shoots of Q.ilex and Q. suber, respectively (Table 2), but differences were notstatistically significant, either between trees or species or sites.

3.4. Xylem anatomy

Current-year shoots of both species have solitary vessels withcircular to elliptical cross sections, arranged in a diagonal to radialpattern (Fig. 4). The apparent ring limit visible in Fig. 4a corre-sponds to a pause in growth observed between the spring andsummer shoot elongation periods. This growing pattern is fre-quently observed in Q. suber trees at the Lezirias site but not at Mi-tra (Pinto et al., 2011). No occluded vessels were observed in any ofthe species. Dmean values of shoots were compared between treeswithin species/site: differences were not statistically significant.Dmean values of shoot xylem for each species or site were25.28 ± 8.1 lm and 22.21 ± 7.4 lm for Q. suber, at Lezirias and Mi-tra, respectively, and 25.61 ± 8.3 lm for Q. ilex at Mitra (Table 3).Differences between these values were only statistically significantfor Q. suber shoots between sites (P-value = 0.0383), with largervessels at the more mesic site (Lezirias).

Deviations between Dmean and Dhw were higher for shoots of Q.suber from Lezirias, reflecting a wider range of conduit diameters.D95 was always higher (14–16%) than Dmean (Table 3), irrespectiveof the species. Vessel diameter distribution displayed a unimodalpattern, particularly for Q. suber (Fig. 5).

Mean vessel density and conductive area were similar betweenspecies and sites (Table 3).

As expected, maximum measured specific hydraulic conductiv-ity (Ks) was lower than the maximum theoretical hydraulic con-ductivity (Ks,t, estimated from vessel anatomy). The ratiobetween mean values of Ks and Ks,t, an estimate of xylem conduct-ing efficiency, was similar for shoots of both species and sites: 57%for Q. ilex and 52% and 59% for Q. suber shoots, at Lezirias and Mitra,respectively. About 90% of the total theoretical hydraulic conduc-tivity was sustained by vessels with diameters wider than 20 lmin both species.

3.5. ‘‘In situ’’ xylem water potentials and minimum safety marginsabove embolism thresholds

The differences between leaf and shoot xylem water potentialsduring summer sunny days were more pronounced at Lezirias thanat Mitra, due to the higher Wl,pd at Lezirias (Fig. 6). Daily minimumWl for Q. suber was approximately �3.0 MPa at both sites (Fig. 6a

d Mitra sites. Statistical t-student tests were performed for the fitted parameters a andspecies/site. Data on Ks were tested through Analyses of Variance using linear mixedAt least 110 measurements were used per species/site. Differences in Ks between treesand Wxyl,99PLC (estimated parameters).

Q. ilex

Mitra(C) (A and B) (B and C)

�3.21 (1.3) ns ns0.38 (0.02) ns ns1.38 (0.42) ns ns�6.10�15.30

Xylem water potential (Ψ xyl , MPa)

0

20

40

60

80

100

-10-8-6-4-20 -10-8-6-4-20

Loss

in h

ydra

ulic

con

duct

ivity

(PLC

,%)

-10-8-6-4-20

(c)(b)(a)

xelisucreuQrebussucreuQ

artiMartiMsairizeL

R2 = 0.69Ψxyl,50PLC = -2.89 MPa

Ψxyl,50PLC = -3.03 MPa

R2 = 0.87Ψxyl,50PLC = -2.94 MPa

Ψxyl,50PLC = -3.08 MPa

R2 = 0.94Ψxyl,50PLC = -3.21 MPa

Ψxyl,50PLC = -3.35 MPa

Fig. 3. Vulnerability curves (VCs) to drought-induced embolism of current-year shoots: Q. suber (circles) from (a) Lezirias site and (b) Mitra site; Q. ilex (c) (triangles) fromMitra site. VCs were adjusted using the original Pammenter and Vander Willigen equation (black lines, Eq. (1)) and a modified version to fit a non-zero ‘‘native’’ level ofembolism (grey lines, Eq. (2)). Each data point represents the average of 3–6 repeated PLC measurements at a given Wxyl for each sampled branch. Error bars are standarderrors of the repeated measurements. Data were collected from four trees per species/site. The xylem tension causing 50% loss in hydraulic conductivity (Wxyl,50PLC) is given asinserts (black: original version; grey: modified version).

Fig. 4. Microscope images of shoot xylem of Q. suber from the Lezirias (a) and Mitra (b) sites and Q. ilex (c) from Mitra site. Scale bars represent 200 lm.

6 C.A. Pinto et al. / Forest Ecology and Management 285 (2012) 1–10

and b), similar to the absolute lowest seasonal Wl,md for Q. suber(Fig. 2a and b). For Q. ilex at Mitra, daily minimum Wl was�3.2 MPa (Fig. 6c), also within the range of the observed lower val-ues of seasonal Wl,md (Fig. 2c). Daily minimum Wxyl of shoots were�1.5 and �2.4 MPa for Q. suber at Lezirias and Mitra, respectively,and �2.3 MPa for Q. ilex (Fig. 6).

Based on the relationships of Fig. 7 (Eq. (8)), we could predictWxyl of shoots from the long-term Wl,md and Wl,pd data (Fig. 2), bothfor Q. suber (Lezirias and Mitra sites) and Q. ilex (Mitra site). For thewhole study period, predicted absolute minimum Wxyl of Q. subershoots was �1.5 MPa at Lezirias and �2.5 MPa at Mitra. For Q. ilexshoots, predicted absolute minimum Wxyl was �2.4 MPa (Mitra).Entering these values in the derived vulnerability curves (Table 2,Fig. 3), we evaluated the most critical embolism conditions experi-enced by shoots under field conditions: PLC values of 42.4% in Q.ilex at Mitra, and of 44.9% and 33.3% in Q. suber at Mitra and Lezi-rias, respectively. Even in the most critical conditions, the expected

in situ maximum loss of conductivity in shoots was below 50%. Theestimated safety margins, above Wxyl,50PLC, were 1.47 and 0.49 MPafor Q. suber shoots at Lezirias and Mitra, respectively, and 0.86 MPafor Q. ilex at Mitra.

4. Discussion

Distinct seasonal patterns of predawn leaf water potential wereobserved at the study sites (Lezirias and Mitra). At Mitra trees suf-fered a considerable water stress during the summer (Fig. 2),whereas at Lezirias Q. suber trees always maintained a favourablewater status. Although rainfall amount and distribution variedwithin and between years at both sites (Fig. 1), a similar seasonalpattern of Wl,pd was observed between years at each site, reflectingthe water availability near the root systems (soil and groundwa-ter). At the end of the summer, when topsoil dries out, the watertable depth was similar at both sites (see Subsection 3.1. in

-3.5

-3.0

-2.5

-2.0

-1.5

-1.0

-0.5

0.0

Q. suber - Mitra: 08-Aug

-3.5

-3.0

-2.5

-2.0

-1.5

-1.0

-0.5

0.0

Q. suber - Lezirias: 22-Jul

00:00

Q. ilex - Mitra: 07-Aug

Ψxy

l; Ψ

l(MPa

)

(b)

(a)

(c)

0.0

00:00 06:00 12:00 18:00

00:00 06:00 12:00 18:00

06:00 12:00 18:00

Q. ilex - Mitra: 07-Aug(c)

Table 3Mean (SD) of anatomical parameters of current-year shoots of Q. suber and Q. ilex,from the Lezirias and Mitra sites (three segments per tree from four trees per species/site). Measured data (Dmean, vessel density and conductive area) were tested throughAnalyses of Variance using linear mixed models considering species/sites as fixedeffects and tree within species as a random factor. Differences between trees withinspecies/site were not statistically significant. Statistics are not given for D95 and Dhw

(estimated parameters).

Parameters Q. suber Q. ilex

Lezirias Mitra Mitra(A) (B) (C) (A and B) (B and C)

Dmean (lm) 25.28(8.1)

22.21(7.4)

25.61(8.3)

a ns

Vessel density(vessel mm�2)

126.30(43.4)

162.44(58.4)

135.04(52.0)

ns ns

Conductivearea (%)

6.99(2.4)

6.99(1.9)

7.68(1.9)

ns ns

D95 (lm) 29.23(4.4)

25.72(4.2)

29.11(5.0)

Dhw (lm) 34.64(6.2)

30.28(4.6)

32.60(5.0)

ns: non-significant.a P < 0.05.

C.A. Pinto et al. / Forest Ecology and Management 285 (2012) 1–10 7

Results). Hence, differences in tree water uptake mainly resultedfrom differences in site lithology. In fact, the sandy nature of thedeep soil at Lezirias did not restrict roots from freely tapping deepwater sources (groundwater). At Mitra, the harder nature of thebedrock (granite, 1 m below soil surface) somehow constrainedthe groundwater uptake by roots. The absolute seasonal minimumWl,md in Q. suber was similar at both sites (approximately�3.0 MPa), regardless the variation in Wl,pd. Hence, minimum leafwater potential seems to have been regulated at an almost con-stant species-specific minimum threshold, by a tight stomatalcontrol.

In respect to xylem vulnerability to embolism, the fitted VCs(Eq. (1)), showed that the 50% loss in hydraulic conductivity of cur-

0

5

10

15

20

25

30

0

5

10

15

20

25

30

35

Vess

el fr

eque

ncy

(% o

f tot

al)

0

5

10

15

20

25

30

0

5

10

15

20

25

30

35(a) Q. suber

(b) Q. ilex

Cla

ss K

h,t(%

of t

otal

)

Vessel diameter: 5-µm classes

Fig. 5. Frequency distribution of the vessel diameter classes (5-lm classes) (bars)and their relative contribution to the total theoretical conductivity (Kh,t) (lines) inshoots of (a) Q. suber from the Lezirias (open bars, full line) and Mitra (closed bars,dashed line) sites and (b) Q. ilex from Mitra site (closed bars, dashed line). Threeshoot segments per tree were sampled from four trees per species/site. Error barsare standard errors.

-3.5

-3.0

-2.5

-2.0

-1.5

-1.0

-0.5

Fig. 6. Daily variation of xylem (open symbols) and leaf (closed symbols) waterpotential in Q. suber (a and b) and Q. ilex (c) measured in typical sunny summerdays, at the Lezirias and Mitra sites. Three covered and three uncovered leaves pertree were measured in four trees per species/site. Error bars are standard errors.

rent-year shoots (Wxyl,50PLC) occurred at xylem water potentials of�2.89 and �2.94 MPa for Q. suber (Lezirias and Mitra, respectively)and �3.21 MPa for Q. ilex (Mitra) (Table 2, Fig. 3). Although thecurves showed some ‘‘native’’ embolism (PLCnat), Wxyl,50PLC doesnot seem to depend on considering or not PLCnat on the formulationof the fitting equation (Eqs. (1) and (2)) (Fig. 3). The observed PLCnat

may reflect the in situ level of cavitation or may result from an arte-fact related to air entry at branch base induced by the initial cut (ifmaximum vessel length exceeds total branch length). The fact thatsimilar PLCnat values have been found in Q. ilex shoots (over 20% inMartínez-Vilalta et al. (2002), and over 10% in Limousin et al.(2010)) and in other Mediterranean oaks (Iovi et al., 2009) maysuggest that PLCnat is real and not an experimental artefact.

Xylem vulnerability to drought-induced embolism was quitesimilar (differences not statistically significant) both between spe-cies (Q. suber and Q. ilex, at Mitra) or sites (Q. suber at Lezirias and

y = 0.5038x + 0.504R² = 0.73

0.0

0.5

1.0

1.5

2.0

2.5

3.0

0.0 2.0 4.0

y = 0.4373x + 0.4971R² = 0.91

2.0

4.0

6.0

8.0

10.0

12.0

0.00.0 10.0 20.0

Ψxy

l / Ψ

l,pd

(MPa

)

(a) Q. suber (b) Q. ilex

Ψl / Ψl,pd (MPa)

Fig. 7. Relationships between Wxyl=Wl;pd and Wl=Wl;pd for Q. suber (a) and Q. ilex (b).

8 C.A. Pinto et al. / Forest Ecology and Management 285 (2012) 1–10

Mitra) (Table 2, Fig. 3). It has been reported to be a species-specifictrait, static across sites and water availability conditions (Iovi et al.,2009; Martínez-Vilalta et al., 2009; Matzner et al., 2001). However,our data suggest the existence of a between-site difference in Dmean

of Q. suber shoots (statistically significant), with wider vessels atthe more mesic site (Lezirias) (Table 3). In the same species, widevessels may be expected to have a larger amount of pit area and agreater probability of large leaky pit pores in the inter-vessel pitmembranes (Jansen et al., 2009). The higher water stress at Mitramay have imposed safety constraints, reducing vessel diameterwhilst maintaining Ks, vessel density and conductive area (Table 2,Table 3), i.e. narrower and safer conduits where water potentialsare more negative (Sperry et al., 2008).

Mediterranean evergreen oaks, such as Q. suber and Q. ilex, arereported to be much less vulnerable to embolism than deciduousoaks from temperate climates with ring-porous wood (Tyree andCochard, 1996). Our findings support these conclusions: Wxyl,50PLC

values for current-year shoots of Q. ilex and Q. suber (�3.2 and�2.9 MPa, or �3.4 and �3.1 MPa if PLCnat is considered, respec-tively) (Table 2, Fig. 3) are more negative than those found for tem-perate climate oaks in North America (varying from �0.50 to�1.89 MPa) (Maherali et al., 2006; Matzner et al., 2001). A widerange of Wxyl,50PLC values are reported for Q. ilex: �2.0 MPa (Martí-nez-Vilalta et al., 2002), �3.1 MPa (Lo Gullo and Salleo, 1993),�3.5 MPa (Tognetti et al., 1998), �3.6 MPa (Limousin et al.,2010), �3.2 MPa (Gartner et al., 2003), �3.0 MPa (Q. ilex subsp. ilex)and �5.6 MPa (Q. ilex subsp. ballota) (Corcuera et al., 2004). Thesedifferences may be ascribed to: (a) different Q. ilex subspecies; (b)the use of seedlings or shoots of adult trees; (c) the use of differentmeasuring techniques (dehydration vs. air injection). Vaz et al.(2012) measured a Wxyl,50PLC of �5.0 MPa for shoots/stems of corkoak seedlings, in the lowest range of the values observed for Q. ilex.Irrespective of the fitted equations (see Fig. 3), our estimates ofWxyl,50PLC for Q. ilex and Q. suber shoots are in the same range ofthose published for adult Q. ilex trees (e.g. Limousin et al., 2010;Tognetti et al., 1998), but are less negative than those reportedby Tyree and Cochard (1996) for both species. Interestingly, thelowest recorded Wxyl,50PLC values for both Q. ilex and Q. suber wereobserved in seedlings (Corcuera et al., 2004; Vaz et al., 2012). Sinceseedlings have not yet a deep root system, they probably need tobe more drought-tolerant than adult trees, i.e. less vulnerable toembolism. A lower vulnerability to cavitation in seedlings than inadult trees has been reported for Q. ilex (Lo Gullo and Salleo,1993), Q. wislizenii (Matzner et al., 2001), Eucalyptus crebra andEucalyptus xanthoclada (Rice et al., 2004). This may impact on seed-ling survival and natural regeneration. Adult trees with fully devel-

oped roots may access deep water sources buffering the seasonalwater deficits (Canadell et al., 1996; David et al., 2007).

Wood anatomy of shoots of Q. ilex and Q. suber showed solitary,circular to elliptical vessels arranged in a diagonal to radial pattern(Fig. 4), which is consistent to that reported in other studies for thesame species (Corcuera et al., 2004; Leal et al., 2008). Vessel diam-eter frequency distribution approached a unimodal pattern (Fig. 5)in shoots of both species (more evident in Q. suber), suggesting adiffuse-type porosity, as reported by Gartner et al. (2003), Lealet al. (2008) and Villar-Salvador et al. (1997). This is also corrobo-rated by the difference between Dmean and D95 for shoots of bothspecies, 14–16%, indicating a small difference between the size ofearly and latewood vessels (Tyree et al., 1994).

In Ks,t calculations all vessels were assumed to be functional. Asexpected, measured Ks was lower than Ks,t since the theoreticalconductivity was calculated neglecting non-functional vesselsand the resistance between vessels (Sellin et al., 2008). Xylem con-ducting efficiency (Ks/Ks,t) of the two oaks was similar, and valuesagree well with those previously published for Q. ilex (Limousinet al., 2010). This is a further indication that PLCnat is probablynot an experimental artefact and that some vessels are not func-tional anymore.

According to Hacke et al. (2000) the resistance to cavitationmay adjust to the amount of drought experienced in situ. How-ever, the non-significant difference in xylem vulnerability toembolism found between Q. ilex at the Mitra site and Q. suberat the Mitra and Lezirias sites suggests that this hydraulic traitdoes not explain per se the current geographical distribution ofthe two species: Q. ilex prevailing in more xeric areas comparedto Q. suber. Additional mechanisms of drought tolerance/avoid-ance may probably be more relevant. For instance, rooting depthmay play an important role since it is known that Mediterraneanevergreen oaks rely on deep roots to cope with the summerdrought (Canadell et al., 1996; Maherali et al., 2004). A previousstudy on sap flow and predawn leaf water potential of co-occur-ring adult Q. suber and Q. ilex trees at the same site (Mitra)showed that Q. ilex trees were able to maintain a higher waterstatus and transpiration during the summer drought comparedto Q. suber (David et al., 2007). This was ascribed to a deeperand more efficient root system in Q. ilex.

Aiming at the evaluation of the in situ safety margins of shootsabove cavitation thresholds, we estimated the long-term patternsof variation of xylem water potential (Fig. 7). Unlike Wl,md, the min-imum value of xylem water potential in shoots does not seem to bemaintained to a constant level across sites but depends both onstomatal regulation and on Wl,pd (Figs. 6 and 7). Wxyl is usually con-

C.A. Pinto et al. / Forest Ecology and Management 285 (2012) 1–10 9

sidered a better estimator of plant water status than Wl,md (Chonéet al., 2001).

The comparison between the predicted minimum Wxyl underfield conditions and Wxyl,50PLC, Wxyl,75PLC and Wxyl,99PLC (Table 2,Fig. 3) suggests that shoots of both species live, at both sites, witha large hydraulic safety margin, above these embolism thresholds.The minimum safety margins observed in shoots of both speciesabove the Wxyl,50PLC were similar at Mitra, although slightly higherfor Q. ilex. The hydraulic safety margins were higher in Q. subershoots at the Lezirias site due to the more favourable tree waterstatus. Wide hydraulic safety margins are a likely protection mech-anism, preventing embolism from occurring in organs where itmay be more damaging (like stems) or where its reversal is lessefficient (like shoots) (Nardini and Salleo, 2000; Pockman andSperry, 2000). Martínez-Vilalta et al. (2002) have observed in Q. ilexan in situ minimum Wxyl in shoots close to Wxyl,75PLC. Lopez et al.(2005) found that, under field conditions, several tropical treesusually lived with minimum shoot Wxyl between Wxyl,50PLC andWxyl,75PLC. Therefore, at our sites, the studied oak trees seem verywell adapted, operating with larger safety margins than those usu-ally reported.

We believe that our results are relevant to the understanding ofsurvival, growth, and functional behaviour of evergreen oaks inMediterranean-type climates, subject to recurrent/seasonaldrought conditions.

Acknowledgements

This work was supported by the Portuguese Foundation for Sci-ence and Technology (FCT) (projects POCI-PTDC/AGR/59152/2004).Clara A. Pinto was funded by a doctoral grant from FCT (SFRH/BD/46479/2008). We thank Rui Alves and Vitor Barros (Companhia dasLezirias) for providing site facilities, Joaquim Mendes (ISA) forassistance in the field work, and Cristiana Duarte (FLOR/IICT) forhelp in the anatomy work. We also thank the two anonymousReviewers for the helpful and constructive comments on themanuscript.

References

Bugalho, M.N., Caldeira, M.C., Pereira, J.S., Aronson, J., Pausas, J.G., 2011.Mediterranean cork oak savannas require human use to sustain biodiversityand ecosystem services. Front. Ecol. Environ. 9 (5), 278–286.

Canadell, J., Jackson, R.B., Ehleringer, J.R., Mooney, H.A., Sala, O.E., Schulze, E.D.,1996. Maximum rooting depth of vegetation types at the global scale. Oecologia108, 583–595.

Choat, B., Drayton, W.M., Brodersen, C., Matthews, M.A., Shackel, K.A., Wada, H.,McElrone, A.J., 2010. Measurement of vulnerability to water stress-inducedcavitation in grapevine: a comparison of four techniques applied to long-vesseled species. Plant, Cell Environ. 33, 1502–1512.

Choné, X., van Leeuwen, C., Dubourdieu, D., Gaudillère, J.P., 2001. Stem waterpotential is a sensitive indicator of grapevine water status. Ann. Bot. 87, 477–483.

Cochard, H., Damour, G., Bodet, C., Tharwat, I., Poirier, M., Améglio, T., 2005.Evaluation of a new centrifuge technique for rapid generation of xylemvulnerability curves. Physiol. Plant. 124, 410–418.

Corcuera, L., Camarero, J.J., Gil-Pelegrín, E., 2004. Effects of a severe drought onQuercus ilex radial growth and xylem anatomy. Trees – Struct. Funct. 18, 83–92.

David, T.S., Henriques, M.O., Kurz-Besson, C., Nunes, J., Valente, F., Vaz, M., Pereira,J.S., Siegwolf, R., Chaves, M.M., Gazarini, L.C., David, J.S., 2007. Water strategiesin two co-occurring Mediterranean evergreen oaks: surviving the summerdrought. Tree Physiol. 27, 793–803.

FAO, 1988. FAO/UNESCO Soil Map of the World. Revised Legend, with Corrections.World Soil Resources Report 60. FAO, Rome (Reprinted as Technical Paper 20,ISRIC, Wageningen, 1994).

Gartner, B.L., Roy, J., Huc, R., 2003. Effects of tension wood on specific conductivityand vulnerability to embolism of Quercus ilex seedlings grown at twoatmospheric CO2 concentrations. Tree Physiol. 23, 387–395.

Hacke, U.G., Sperry, J.S., Pittermann, J., 2000. Drought experience and cavitation insix shrubs from the Great Basin, Utah. Basic Appl. Ecol. 1, 31–41.

INMG, 1991a. O clima de Portugal. Normais climatológicas da Região Alentejo eAlgarve, correspondentes a 1951–1980, Fascículo XLIX, vol. 4-4ª região.Instituto Nacional de Meteorologia e Geofísica, Lisboa.

INMG, 1991b. O clima de Portugal. Normais climatológicas da Região Ribatejo eOeste, correspondentes a 1951–1980, Fascículo XLIX, vol. 2-2ª região. InstitutoNacional de Meteorologia e Geofísica, Lisboa.

Iovi, K., Kolovou, C., Kyparissis, A., 2009. An ecophysiological approach of hydraulicperformance for nine Mediterranean species. Tree Physiol. 29, 889–900.

Jansen, S., Choat, B., Pletsers, A., 2009. Morphological variation of intervessel pitmembranes and implications to xylem function in angiosperms. Am. J. Bot. 96(2), 409–419.

Leal, S., Nunes, E., Pereira, H., 2008. Cork oak (Quercus suber L.) wood growth andvessel characteristics variations in relation to climate and cork harvesting. Eur.J. Forest Res. 127, 33–41.

Limousin, J.M., Longepierre, D., Huc, R., Rambal, S., 2010. Change in hydraulic traitsof Mediterranean Quercus ilex subjected to long-term throughfall exclusion.Tree Physiol. 30, 1026–1036.

Lloret, F., Siscart, D., Dalmases, C., 2004. Canopy recovery after drought dieback inholm-oak Mediterranean forests of Catalonia (NE Spain). Global Change Biol. 10,2092–2099.

Lo Gullo, M.A., Salleo, S., 1993. Different vulnerabilities of Quercus ilex L. to freeze-and summer drought-induced xylem embolism: an ecological interpretation.Plant, Cell Environ. 16, 511–519.

Lo Gullo, M.A., Salleo, S., Rosso, R., Trifilò, P., 2003. Drought resistance of 2-year-oldsaplings of Mediterranean forest trees in the field: relations between waterrelations, hydraulics and productivity. Plant Soil 250, 259–272.

Lopez, O.R., Kursar, T.A., Cochard, H., Tyree, M.T., 2005. Interspecific variation inxylem vulnerability to cavitation among tropical tree and shrub species. TreePhysiol. 25, 1553–1562.

Maherali, H., Pockman, W.T., Jackson, R.B., 2004. Adaptive variation in thevulnerability of woody plants to xylem cavitation. Ecology 85 (8), 2184–2199.

Maherali, H., Moura, C.F., Caldeira, M.C., Wilson, C.J., Jackson, R.B., 2006. Functionalcoordination between leaf gas exchange and vulnerability to xylem cavitationin temperate forest trees. Plant, Cell Environ. 29, 571–583.

Martínez-Vilalta, J., Prat, E., Oliveras, I., Piñol, J., 2002. Xylem hydraulic properties ofroots and stems of nine Mediterranean woody species. Oecologia 133, 19–29.

Martínez-Vilalta, J., Cochard, H., Mencuccini, M., Sterck, F., Herrero, A., Korhonen,J.F.J., Llorens, P., Nikinmaa, E., Nolé, A., Poyatos, R., Ripullone, F., Sass-Klaassen,U., Zweifel, R., 2009. Hydraulic adjustment of Scots pine across Europe. NewPhytol. 184, 353–364.

Matzner, S.L., Rice, K.J., Richards, J.H., 2001. Intra-specific variation in xylemcavitation in interior live oak (Quercus wislizenii). J. Exp. Bot. 52 (357), 783–789.

McDowell, N.G., Beerling, D.J., Breshears, D.D., Fisher, R.A., Raffa, K.F., Stitt, M., 2011.The interdependence of mechanisms underlying climate-driven vegetationmortality. Trends Ecol. Evol. 26 (10), 523–532.

Miranda, P., Coelho, F.E.S., Tomé, A.R., Valente, M.A., 2002. 20th Century Portugueseclimate and climate change scenarios. In: Santos, F.D., Forbes, K., Moita, R.(Eds.), Climate Change in Portugal. Scenarios, Impacts and AdaptationMeasures. Gradiva, Lisboa, pp. 25–83.

Nardini, A., Salleo, S., 2000. Limitation of stomatal conductance by hydraulic traits:sensing or preventing xylem cavitation? Trees – Struct. Funct. 15, 14–24.

Ogle, K., Barber, J.J., Willson, C., Thompson, B., 2009. Hierarchical statisticalmodeling of xylem vulnerability to cavitation. New Phytol. 182, 541–554.

Pammenter, N.W., Vander Willigen, C., 1998. A mathematical and statistical analysisof the curves illustrating vulnerability of xylem cavitation. Tree Physiol. 18,589–593.

Pinto, C.A., Henriques, M.O., Figueiredo, J.P., David, J.S., Abreu, F.G., Pereira, J.S.,Correia, I., David, T.S., 2011. Phenology and growth dynamics in Mediterraneanevergreen oaks: effects of environmental conditions and water relations. For.Ecol. Manage. 262, 500–508.

Pockman, W.T., Sperry, J.S., 2000. Vulnerability to xylem cavitation and thedistribution of Sonoran desert vegetation. Am. J. Bot. 87 (9), 1287–1299.

Rice, K.J., Matzner, S.L., Byer, W., Brown, J.R., 2004. Patterns of tree dieback inQueensland, Australia: the importance of drought stress and the role ofresistance to cavitation. Oecologia 139, 190–198.

Ryan, M.G., 2011. Tree responses to drought. Tree Physiol. 31, 237–239.Scholander, P.F., Hammel, H.T., Bradstreet, E.D., Hemmingsen, E.A., 1965. Sap

pressure in vascular plants. Science 148, 339–346.Sellin, A., Rohejärv, A., Rahi, M., 2008. Distribution of vessel size, vessel density and

xylem conducting efficiency within a crown of silver birch (Betula pendula).Trees – Struct. Funct. 22, 205–216.

Snedecor, G.W., Cochran, W.G., 1978. Statistical Methods, sixth ed. The Iowa StateUniversity Press, Ames.

Sperry, J.S., 1986. Relationship of xylem embolism to xylem pressure potential,stomatal closure, and shoot morphology in palm Rhapis excelsa. Plant Physiol.80, 110–116.

Sperry, J.S., Donnelly, J.R., Tyree, M.T., 1988. A method for measuring hydraulicconductivity and embolism in xylem. Plant, Cell Environ. 11, 35–40.

Sperry, J.S., Nichols, K.L., Sullivan, J.E.M., Eastlack, S.E., 1994. Xylem embolism inring-porous, diffuse-porous and coniferous trees of northern Utah and interiorAlaska. Ecology 75 (6), 1736–1752.

Sperry, J.S., Meinzer, F.C., McCulloh, K.A., 2008. Safety and efficiency conflicts inhydraulic architecture: scaling from tissues to trees. Plant, Cell Environ. 31,632–645.

Sperry, J.S., Christman, M.A., Torres-Ruiz, J.M., Taneda, H., Smith, D.D., 2012.Vulnerability curves by centrifugation: is there an open vessel artefact, and are‘r’ shaped curves necessarily invalid? Plant, Cell Environ. 35, 601–610.

Tognetti, R., Longobucco, A., Raschi, A., 1998. Vulnerability of xylem to embolism inrelation to plant hydraulic resistance in Quercus pubescens and Quercus ilex co-

10 C.A. Pinto et al. / Forest Ecology and Management 285 (2012) 1–10

occurring in a Mediterranean coppice stand in central Italy. New Phytol. 139,437–447.

Tyree, M.T., Cochard, H., 1996. Summer and winter embolism in oak – impact onwater relations. Ann Sci Forest 53, 173–180.

Tyree, M.T., Sperry, J.S., 1988. Do woody plants operate near the point ofcatastrophic xylem dysfunction caused by dynamic water stress? PlantPhysiol. 88, 574–580.

Tyree, M.T., Zimmermann, M.H., 2002. Xylem structure and the ascent of sap. In:Tyree, M.T. (Ed.), Springer Series in Wood Science, second ed. Springer-Verlag,Berlin.

Tyree, M.T., Davis, S.D., Cochard, H., 1994. Biophysical perspectives of xylemevolution: is there a trade-off of hydraulic efficiency for vulnerability todysfunction? IAWA J. 15 (4), 335–360.

van Mantgem, P.J., Stephenson, N.L., Byrne, J.C., Daniels, L.D., Franklin, J.F., Fule, P.Z.,Harmon, M.E., Larson, A.J., Smith, J.M., Taylor, A.H., Veblen, T.T., 2009.

Widespread increase of tree mortality rates in the Western United States.Science 323, 521–524.

Vaz, M., Cochard, H., Gazarini, L., Graça, J., Chaves, M.M., Pereira, J.S., in press. Corkoak (Quercus suber L.) seedlings acclimate to elevated CO2 and water stress:photosynthesis, growth, wood anatomy and hydraulic conductivity. Trees –Struct. Funct. doi: http://dx.doi.org/10.1007/s00468-012-0691-x.

Villar-Salvador, P., Castro-Díez, P., Pérez-Rontomé, C., Montserrat-Martí, G., 1997.Stem xylem features in three Quercus (Fagaceae) species along a climaticgradient in NE Spain. Trees – Struct. Funct. 12, 90–96.

Wullschleger, S.D., Meinzer, F.C., Vertessy, R.A., 1998. A review of whole-plant wateruse studies in trees. Tree Physiol. 18, 499–512.

Zar, J.H., 1996. Biostatistical Analysis, third ed. Prentice Hall International, NewJersey.

Zimmermann, M.H., Jeje, A.A., 1981. Vessel-length distribution in some Americanwoody plants. Can. J. Bot. 59, 1882–1892.